Hiding in plain sight with vibe coding.

CyberWire Daily

N2K Networks, Inc.

Daily News, Tech News, News, Technology

4.6 • 1K Ratings

🗓️ 14 June 2025

⏱️ 21 minutes

🔗️ Recording | iTunes | RSS

🧾️ Download transcript

Summary

This week, Dave is joined by ⁠Ziv Karliner⁠, ⁠Pillar Security⁠’s Co-Founder and CTO, sharing details on their work on "New Vulnerability in GitHub Copilot and Cursor: How Hackers Can Weaponize Code Agents." Vibe Coding - where developers use AI assistants like GitHub Copilot and Cursor to generate code almost instantly - has become central to how enterprises build software today. But while it’s turbo-charging development, it’s also introducing new and largely unseen cyber threats. The team at Pillar Security identified a novel attack vector, the ⁠"Rules File Backdoor"⁠, which allows attackers to manipulate these platforms into generating malicious code. It represents a new class of supply chain attacks that weaponizes AI itself, where the malicious code suggestions blend seamlessly with legitimate ones, bypassing human review and security tools. The research can be found here: ⁠New Vulnerability in GitHub Copilot and Cursor: How Hackers Can Weaponize Code Agents Learn more about your ad choices. Visit megaphone.fm/adchoices

Transcript

Click on a timestamp to play from that location

0:00.0	You're listening to the Cyberwire Network, powered by N2K.
0:09.7	And now a word from our sponsor, Threat Locker.
0:17.5	Keeping your system secure shouldn't mean constantly reacting to threats. Threat Locker
0:22.7	helps you take a different approach by giving you full control over what software can run in your
0:27.3	environment. If it's not approved, it doesn't run. Simple as that. It's a way to stop ransomware
0:33.2	and other attacks before they start without adding extra complexity to your day. See how Threatlocker
0:39.3	can help you lock down your environment at www.threatlocker.com. Hello everyone and welcome to the CyberWires Research Saturday.
1:01.0	I'm Dave Bittner, and this is our weekly conversation with researchers and analysts
1:06.0	tracking down the threats and vulnerabilities, solving some of the hard problems and protecting ourselves in a rapidly evolving cyberspace.
1:14.6	Thanks for joining us.
1:20.6	Pillar Security, we spent the last year and a half, spent a lot of time with the emerging attack vectors that put
1:29.0	AI-powered applications at risk.
1:32.2	So first of all, we got to learn and get our hands around a new attack vectors such as prompt
1:39.5	injection, indirect injections and all sorts of evasion techniques that turn this attacks to be basically invisible to the human eye and most of the security tools out there.
1:51.7	That's Ziv Karliner, Pillar Securities co-founder and CTO. The research we're discussing today is titled,
1:59.5	New Vulnerability in GitHub Copilot and Cursor,
2:03.1	how hackers can weaponize code agents.
2:13.5	So take that together with the fact that we ourselves are utilizing these amazing coding co-pilots that on their own are utilizing LLM and its base,
2:27.3	got us thinking about how the combination of the new attack vectors and the actual I would say some of the most popular use cases for the
2:39.4	AI powered applications which are coding assistance how this really combines together and
2:49.0	sparked our imagination about what can potentially go wrong.
2:53.4	Well, at the root of this is what you all refer to as the rules file back door. Can you describe that for us?
	...

Transcript will be available on the free plan in 27 days. Upgrade to see the full transcript now.

Previous episode

Disclaimer: The podcast and artwork embedded on this page are from N2K Networks, Inc., and are the property of its owner and not affiliated with or endorsed by Tapesearch.

Generated transcripts are the property of N2K Networks, Inc. and are distributed freely under the Fair Use doctrine. Transcripts generated by Tapesearch are not guaranteed to be accurate.